Jump to content

Draft:DiffusionData Limited

From Wikipedia, the free encyclopedia


DiffusionData Limited is a British software company headquartered in Reading, Berkshire, that develops Diffusion, a secure, real-time pub/sub server built for mobile, web, and AI applications.[1][2] DiffusionData specialises in publish–subscribe solutions for event-driven applications in sectors including financial services and e‑gaming.[2][3]

The company's flagship product, Diffusion, is described in independent technical commentary as a real-time data management platform built around topics and publish–subscribe streams that consume, enrich and distribute event data in real time via a publish–subscribe model across on‑premise, cloud and hybrid environments with fine-grained access control.[4][5][6]

The company’s platform is used in production by major e‑gaming operators; industry coverage reports that Push Technology (now DiffusionData) has powered over £5.2 billion in annual bets in 2017 for brands including William Hill, 888 Holdings, Betsson, Paddy Power, Racing Post and BetDaq.[7] In 2020, Betsson Group selected Diffusion for a global, cloud-based sportsbook rollout to deliver real-time betting odds across multiple brands and geographic markets.[8][9]

In 2025, DiffusionData announced the Diffusion MCP Server, an implementation of the Model Context Protocol designed to act as an AI gateway to its real-time data platform, allowing MCP‑compatible assistants to explore topics, sessions and metrics and perform operations in natural language.[10][11]

History

[edit]

DiffusionData was incorporated on 23 February 2006 in England and Wales under the name CIED Ltd., later changing its registered name to Push Technology Limited in 2008, then to DiffusionData Limited in 2022.[1] The company received venture funding while trading as Push Technology, and a £1 million funding package from the North East Development Capital Fund (NEDCF) in 2023 to open an engineering hub in Newcastle and support product development.[6][12][13] An external profile notes that the business, now operating as DiffusionData, positions itself as addressing the “data in motion” market with its intelligent data platform.[2]

Technology

[edit]

Independent analysts describe Diffusion as a real-time data management and distribution platform built around topics and publish–subscribe streams that consume, enrich and distribute event data between back-end systems and client applications.[4][5] DiffusionData states that the platform is designed to support publish–subscribe architectures across on‑premise, cloud and hybrid environments.[14]

Key aspects of Diffusion’s data model include:[15][16]

  • Clients subscribe to named topics, and the server pushes updates only when those topics change, rather than relying on client polling.[15]
  • Topics are uniquely identified data streams that can represent structured or scalar values and can be transformed or filtered before delivery.[15][4]

Diffusion exposes several topic types:[17][18][19]

  • JSON topics, which store JSON values and can be delivered using structural deltas in a binary encoding for efficiency.[18]
  • Binary topics, which hold arbitrary binary payloads and support delta streaming of binary data.[20]
  • String, double and int64 topics for simple scalar values.[19]
  • RecordV2 topics, which manage structured records with per‑field deltas.[19]
  • Time-series topics, which store ordered sequences of events with configurable retention policies.[19]

Topic behaviour is controlled by properties such as:[17][19]

  • conflation policies (for example OFF, CONFLATE, ALWAYS) that determine whether queued updates are merged or dropped for slow consumers;
  • compression settings that control whether queued updates are compressed in server queues and on the wire;
  • per-topic update throttling and constraints used by the server’s flow-control system.

According to DiffusionData, the platform’s transport layer uses a proprietary binary protocol over WebSockets with delta streaming and optional compression:[14][16]

  • only changes between successive values are transmitted, rather than full payloads;
  • delta streaming and compression are intended to reduce network bandwidth and end‑to‑end latency for suitable workloads;
  • marketing material claims “over 90%” reductions in bandwidth and “90% fewer server resources” for data distribution in some use cases, although independent sources do not provide separate quantified measurements.[14][21]

Operational behaviour is supported by additional mechanisms:[22]

  • per-session queues and conflation are used to adapt to congestion and slow or disconnected clients while keeping subscribers updated with the latest values;[22]
  • topic and session metrics, plus monitoring and alerting features, expose data about throughput and latency for operational teams.[3]

References

[edit]

[1] [2] [3] [4] [5] [6] [7] [8] [9] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] </references>

  1. ^ a b c "DIFFUSIONDATA LIMITED overview". Companies House. UK Government. Retrieved 17 December 2025.
  2. ^ a b c d "How DiffusionData Is Addressing The $300 Billion Data-In-Motion Opportunity". Pulse 2.0. 11 April 2023. Retrieved 17 December 2025.
  3. ^ a b c "Push Technology releases Diffusion 6.0". IoT Now. 19 October 2017. Retrieved 17 December 2025.
  4. ^ a b c d "Diffusion Intelligent Data Platform". Bloor Research. Retrieved 17 December 2025.
  5. ^ a b c "Push Technology releases Diffusion Intelligent Data Platform 6.2". Help Net Security. 13 November 2018. Retrieved 17 December 2025.
  6. ^ a b c "DiffusionData receives £1 million investment from NEDCF". Maven Capital Partners. 6 March 2023. Retrieved 17 December 2025.
  7. ^ a b "UK – Over £5bn eGaming bets processed using Push Technology". G3 Newswire. 19 February 2017. Retrieved 17 December 2025.
  8. ^ a b "Push Technology's Diffusion chosen by Betsson Group for Global Sportsbook Cloud Rollout". Gaming Americas. 29 June 2020. Retrieved 17 December 2025.
  9. ^ a b "Betsson Group chooses Push Technology for cloud-based sportsbook roll-out". EGR Intel. 29 June 2020. Retrieved 17 December 2025.
  10. ^ "DiffusionData Targets Agentic AI in Finance with New MCP Server". A-Team Insight. 7 November 2025. Retrieved 5 January 2026.
  11. ^ "DiffusionData Introduces MCP Server to Maximize the Value of Real-Time Data". Database Trends and Applications. 5 November 2025. Retrieved 5 January 2026.
  12. ^ a b "DiffusionData receives £1million funding from the North East Fund". North East Fund. 9 November 2025. Retrieved 17 December 2025.
  13. ^ a b "Funds and Newcastle Opening for DiffusionData". MrWeb. 8 March 2023. Retrieved 17 December 2025.
  14. ^ a b c d "Diffusion Intelligent Event-Data Platform Datasheet" (PDF). DiffusionData. 2025. Retrieved 17 December 2025.
  15. ^ a b c d "Introducing topics and data". Diffusion Documentation. DiffusionData. 31 December 2024. Retrieved 17 December 2025.
  16. ^ a b c "Push Technology introduces powerful functionality for dynamic data management". IoT Now. 8 May 2018. Retrieved 17 December 2025.
  17. ^ a b c "Properties of topics". Diffusion Documentation. DiffusionData. Retrieved 17 December 2025.
  18. ^ a b c "JSON topics". Diffusion Documentation. DiffusionData. Retrieved 17 December 2025.
  19. ^ a b c d e f "Topics – Python SDK for Diffusion". Diffusion Documentation. DiffusionData. Retrieved 17 December 2025.
  20. ^ a b "Binary topics". Diffusion Documentation. DiffusionData. Retrieved 17 December 2025.
  21. ^ a b "Signal Centre case study" (PDF). DiffusionData. 2021. Retrieved 17 December 2025.
  22. ^ a b c "Diffusion 6.12.1 User Guide". Diffusion Documentation. DiffusionData. Retrieved 5 January 2026.
Retrieved from "https://en.wikipedia.org/w/index.php?title=Draft:DiffusionData_Limited&oldid=1335150892"